Treating a partially ordered typology as nominal leaves its comparative hypotheses untested

Evidence

Because she treats the typology as nominal rather than ordinal she does not test a number of the hypotheses embedded in the partial ordering, but which she gives in her table 1.3. All that the statistics in table 8.2 tell you is, say, that strongman is significantly different from machine. However, the partial ordering also hypothesizes that (1) strongman is significantly more likely to be involved in a militarized dispute than junta or boss and (2) junta and boss are not significantly different from each other. — Gary Goertz, Social Science Concepts and Measurement, 1294